JP2019061501A - Behavior information processing device - Google Patents

Abstract

To identify the behavior of a skilled worker and to support extracting a method that can be reduced to practice by a non-skilled worker.SOLUTION: A behavior information processing device comprises: an information collection section 130 for collecting the information of a plurality of variables in a work environment; a first information retaining section 140 for retaining the information collected by the information collection section 130; a movement information management section 110, an impact information management section 120, and a relationship information acquisition section 160 for identifying the behaviors of a worker in the work environment; a feature detection section 150 for detecting, from the information retained in the first information retaining section 140, information which an event including predetermined features corresponding to the types of information is expressed therein; the relationship information acquisition section 160 for requesting the input of information relating to the relationship between the behaviors identified by the movement information management section 110, the impact information management section 120, and the relationship information acquisition section 160 and the event of the information detected by the feature detection section 150; and a second information retaining section 170 for retaining the information inputted in response to the request by the relationship information acquisition section 160.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an action information processing apparatus.

  In various operations, a skilled worker may perform operations in a procedure different from that of the non-skilled worker, or the skilled worker may take an action that the non-skilled worker does not take. It is required to inherit such an action by a skilled worker which causes an effect different from that of the work by the non-skilled worker as a skill of the skilled worker.

  In Patent Document 1, a measuring unit for grasping a situation where an object to be worked is placed and a measuring unit are controlled, and appropriate information processing is performed on obtained information, and further to an operator. An information processing unit that constructs visual information to be presented, and a display unit that gives the operator visual information superimposed on the target object, and the worker instructs the work to be performed at the work site A support device is disclosed. Further, Patent Document 2 discloses a factor estimation support device that assists in estimating a factor from a result generated in a production system. This factor estimation support device stores in the storage unit the material / environment history data and inspection history data acquired from the production system, and the causal structure data indicating a causal relationship between a plurality of variables, and the final quality abnormality detection unit When it is determined that the final quality characteristic is abnormal, it is determined whether each variable other than the final quality characteristic is abnormal in the variable abnormality detection unit, and the visualization result of the determination result is visualized as the causal structure data. Reflected in the visualized image. In addition, Patent Literature 3 discloses a degree of interest calculation unit that calculates the degree of interest of a person to an object of measurement based on the positional relationship between a person and one or more objects of measurement and the temporal change of the position relationship. And an attribute storage unit storing one or more attributes regarding the measurement target, an attribute value of the attribute, and an interest level of the measurement target as the interest level of the one or more attributes regarding the measurement target, An information processing apparatus is disclosed that includes a correlation calculation unit that calculates a correlation between a gender or an attribute and an interest level of one or more attributes.

JP, 2009-69954, A Patent No. 5186956 JP, 2010-20718, A

  The behavior of a worker at a work place etc. that is not found in other workers may contribute to the improvement of the work effect, but the reason why it is not verbalized or explained for the worker itself (so-called implicit Knowledge). Such an action is not verbalized and described, so it is difficult to reuse (inherit) as a method, unlike the procedure and the like transmitted in a manual or the like.

  An object of the present invention is to support identification of worker's behavior and extraction as a method that can be implemented by other workers.

The present invention according to claim 1 is
Collection means for collecting information on multiple variables in the work environment;
First holding means for holding information collected by the collecting means;
Action specifying means for specifying a worker's action in the work environment;
A detection unit that detects, from the information held by the first holding unit, information in which an event having a predetermined feature according to the type of the information has appeared;
Request means for requesting input of information on the relationship between the action specified by the action specifying means and the event of the information detected by the detection means;
Second holding means for holding information input in response to the request of the request means;
An action information processing apparatus comprising:
The present invention according to claim 2 is
The request means is
Question presenting means for generating and presenting a question;
Answer acceptance means for accepting input of an answer to the presented question;
The behavior information processing apparatus according to claim 1, comprising:
The present invention according to claim 3 is
3. The action information processing apparatus according to claim 2, wherein the question presenting unit determines the type of the question and the presentation order according to the type of the action specified by the action specifying means.
The present invention according to claim 4 is
The action identification means
Operation specifying means for specifying an operator's operation;
Impact identification means for identifying an event that is determined to be an impact of an operation;
The behavior information processing apparatus according to claim 1, comprising:
The present invention according to claim 5 is
5. The action information according to claim 4, wherein the action specifying unit receives an input of information on an action by the worker, analyzes the input information, and specifies contents of the action of the worker. It is a processing device.
The present invention according to claim 6 is
The behavior information processing apparatus according to claim 4, wherein the operation specifying means specifies the content of the operation of the worker based on data of the worker measured by a sensor provided in the work environment. It is.
The present invention according to claim 7 is
The behavior information processing apparatus according to claim 4, wherein the operation specifying unit analyzes an image obtained by photographing the worker to specify the content of the operation of the worker.
The present invention according to claim 8 is
With regard to the information of the variable acquired in the work environment, information presenting an event having a predetermined characteristic according to the type of information is presented, and the relationship with the worker's behavior in the work environment is given to the worker Inquiry means to inquire,
Response receiving means for receiving an input of a response to the query by the query means;
An action information processing apparatus comprising:

According to the invention of claim 1, the relationship between the action of the worker and the event is compared with the configuration in which the action of the worker is not associated with the event detected in the information related to the target of the action or the task. It is possible to make the worker recognize and to extract the worker's action as a method that can be implemented by other workers.
According to the second aspect of the present invention, in comparison with the configuration in which the worker's behavior is not associated with the event detected in the information related to the target of the behavior or the task, the worker's behavior and events and The operator's perception of the relationship between
According to the third aspect of the present invention, the question can be presented flexibly according to the action of the worker, as compared with the configuration in which the question presentation order is fixedly determined.
According to the fourth aspect of the present invention, compared with the configuration in which the action of the worker is not associated with the event detected in the information related to the target of the action or the task, based on the influence of the worker's action and action Thus, the behavior of the worker can be identified, and the worker's perception of the relationship between the identified behavior and the event can be clarified.
According to the fifth aspect of the present invention, it is possible to clarify the worker's recognition of the relationship with the event regarding the action which the worker himself is aware of.
According to the sixth aspect of the present invention, it is possible to clarify the worker's recognition of the relationship with the event regarding the operation which the worker is not aware of, but specified by the sensor value.
According to the seventh aspect of the present invention, it is possible to clarify the worker's recognition of the relationship with the event with respect to the action identified by analyzing the image taken by the worker but not aware of the worker .
According to the invention of claim 8, as compared with the configuration in which the action of the worker is not associated with the event detected in the information related to the target of the action or the task, the action and the event of the worker are interrogated through the question. The worker can be made to recognize the relationship between the above and the worker's action, and it can help extracting the action of the worker as a method that can be performed by other workers.

It is a figure which shows the structure of the action information processing apparatus with which this embodiment is applied. It is a figure which shows the example of object attribute information. It is a figure which shows the example of test value information. It is a figure which shows the example of work relevant information. It is a figure showing the example of hardware constitutions of the computer used as an action information processing device. It is a flowchart which shows operation | movement of an action information processing apparatus. It is a figure which shows the example of a question screen. It is a figure which shows the other example of a question screen.

Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.
<Configuration of Behavior Information Processing Device>
FIG. 1 is a diagram showing the configuration of an action information processing apparatus to which the present embodiment is applied. The behavior information processing apparatus 100 according to the present embodiment includes an operation information management unit 110, an influence information management unit 120, an information collection unit 130, a first information holding unit 140, a feature detection unit 150, and a relationship information acquisition unit. 160 and a second information holding unit 170. In addition, the behavior information processing apparatus 100 is connected to various external devices for acquiring various information related to a product, a work environment, an operation of a worker, and the like.

  The operation information management unit 110 is means for managing information (operation information) related to the operation performed by the worker. The operation information is information for specifying the operator's operation. Therefore, the operation information management unit 110 is an example of the operation specifying unit. The motion information includes, for example, information on a worker who performed the motion, a target object of the motion, a date and time when the motion was performed, and a content of the motion, as the information specifying the motion. The operation information may not include all of the information, or may include other information. Here, the operation of the worker includes not only the operation in the operation but also the operation performed in other than the operation. For example, the operation of adjusting the equipment used in the operation, the operation of changing the arrangement of tools used in the operation, the operation of changing the standing position of the worker himself or the direction of the product to be operated during the operation, and the like are included. Further, the target includes not only the product to be worked but also the equipment and tools used in the work.

  The operation information managed by the operation information management unit 110 is identified by, for example, identification information of an object. The operation information is acquired, for example, by the worker himself operating the input means as the user interface of the behavior information processing apparatus 100 to input. In addition, it is possible to analyze the change of the image and the sensor value which are acquired by the camera and the sensor which are provided in the work place, specify the operation and acquire the operation information. In this case, the analysis of the video and the sensor value may be performed by an analysis server (external server) prepared for analysis processing, and the operation information management unit 110 may acquire and manage the analysis result.

  The influence information management unit 120 is a means for managing information (influence information) related to the influence caused by the operation performed by the worker. The influence information is information for specifying the influence caused by the operation of the worker. Therefore, the influence information management unit 120 is an example of the influence identification unit. The influence information includes, for example, information indicating an event that has occurred to the object, a change in the state of the object, and the like. Further, the influence information is managed in association with the operation information managed by the operation information management unit 110. The influence information managed by the influence information management unit 120 is identified, for example, by the identification information of the object. The influence information is acquired, for example, by the worker himself operating the input means as the user interface of the behavior information processing apparatus 100 to input. The influence information may be acquired by analyzing the change in the image or sensor value acquired by the camera or sensor provided in the work place, and specifying the influence that has occurred on the object. In this case, the analysis of the image and the sensor value may be performed by an analysis server (external server) prepared for analysis processing, and the influence information management unit 120 may acquire and manage the analysis result.

  The information collecting unit 130 is a means for acquiring information on variables related to the object and the work. The information collecting unit 130 acquires object attribute information and test value information as information of a variable related to the object. In addition, the information collection unit 130 acquires work related information and worker operation information as the information of the variable related to the work. The information collecting unit 130 is an example of a collecting unit. These pieces of information are acquired from, for example, a server (external server) such as a management system that manages products, a management system that manages equipment used in work, and a management system that manages workers. Note that, after the relationship information acquisition unit 160 identifies the action of the worker who is the subject of the relationship information acquisition, the information collection unit 130 relates to the identified action among the information related to the object and the work. Information may be selected and collected. The process in which the relationship information acquisition unit 160 specifies the action of the worker who is the target of the relationship information acquisition will be described later.

  The object attribute information is attribute information of each object of the object. Specifically, information such as the date of manufacture of the object, the date of inspection, etc. is included. Further, depending on the type of the object, information such as the size, weight, shape, color, operating speed and movable range of the movable object, and component parts is included.

  FIG. 2 is a diagram showing an example of object attribute information. FIG. 2 shows attribute information on the cylinder of the device used in the operation as an example of the object. In the example shown in FIG. 2, “attribute name” indicates the type of attribute. The cylinder which is the object is identified by the object ID. In the example illustrated in FIG. 2, five types of attributes of “production date”, “final inspection date”, “cylinder diameter”, “operating speed”, and “component part” for the cylinder of the object ID “# 22” The attribute value of is recorded. Further, in the example shown in FIG. 2, with reference to the attribute value of the column of the attribute name "component", the part specified by the ID "X222" is attached to the cylinder of the object ID "# 22" .

  The inspection value information is information on inspection values related to the attributes of each object. The inspection value is an actual measurement value obtained by inspection on the object. The inspection value of the inspection value information is repeatedly acquired as time passes. Therefore, by taking into consideration the inspection value information, it is possible to trace the transition of the attribute of the object with time.

  FIG. 3 is a diagram showing an example of inspection value information. In FIG. 3, inspection values of two types of attributes of the attribute name “operating speed” and the attribute name “operating sound” are recorded with respect to the cylinder of the object ID “# 22”. Test values are acquired once a day, and FIG. 3 shows test values from May 28 to June 2. Referring to FIG. 3, the test value of the operation sound is stable at approximately 57 db (decibel) from May 28 to June 1, but it is 77 db only on June 2 and other test values There is a big gap. In other words, it can be seen that a loud operation noise was suddenly generated on June 2nd.

  Work-related information is attribute information related to work. Specifically, for example, environmental information such as room temperature, information of a worker who is a work subject, and the like are included. Each piece of information in the work related information is identified, for example, by the place and time when the work was performed.

  FIG. 4 is a diagram showing an example of work related information. In the example shown in FIG. 4, "place" specifies the place where the work was performed. The “start date and time” and the “end date and time” specify the time when the work was performed. The "attribute name" and the "value" indicate the type of attribute and the attribute value as attribute information. In the example shown in FIG. 4, regarding two places of "section X" and "section Y", 11 o'clock to 12 o'clock, 12 o'clock to 13 o'clock, 13 o'clock to 14 o'clock, 14 o'clock on June 2, 2017 In each time zone of 15 o'clock, attribute information on the attribute name "room temperature" and the attribute name "worker" is shown.

  The worker operation information is information for specifying a worker's operation. Worker operations include operations in operations and operations other than operations performed during operations. For example, the operation etc. which change the standing position and posture to a subject are also included. Information on such operation can be obtained by analyzing an image captured by a camera provided at a work place, sensor values detected by various sensors, and the like. In addition, when information of an operation is obtained based on a video or a sensor value, a certainty factor of the operation may be added to the obtained information of the operation. The content of the certainty factor of an operation can be considered by, for example, evaluating and distinguishing the case where it is determined that the operation has been reliably performed and the case where the operation may have been performed in several stages. The information on the degree of certainty added to the worker operation information may be, for example, information indicating an evaluation content such as "performed", "perhaps", "possibly performed", etc. In addition, it may be a numerical value representing the stage of evaluation (such as a percentage value with 100% in the case of being reliably performed, or a 10-step numerical value with the case of being reliably performed as the maximum value). Each piece of information in the worker operation information is identified based on, for example, information on the time, place, and worker when the work was performed. An example of the worker operation information is not particularly illustrated.

  The first information holding unit 140 is a holding unit that holds information acquired by the information collecting unit 130. That is, the first information holding unit 140 is an example of the first holding unit. The information stored in the first information storage unit 140 is referred to in the processing by the feature detection unit 150.

  The feature detection unit 150 is processing means for detecting, among the information held in the first information holding unit 140, information in which a specific event has appeared. An event to be detected is an event having a predetermined feature according to the type of information stored in the first information storage unit 140. Specific events can be set variously according to the configuration concerning work, such as information type of variable and attribute value, product of work target, environment of work place, kind of work, setting of management system for product and work, etc. . For example, a sudden change occurs in the transition of the value of the variable or the attribute value with the passage of time, that the value of the variable or the attribute value exceeds a predetermined threshold, the number of occurrences or occurrence of such change The frequency or the like may be an event of the detection condition. The feature detection unit 150 is an example of a detection unit.

  The relationship information acquisition unit 160 is means for acquiring information on the relationship between the operation managed by the operation information management unit 110 and the event of the information detected by the feature detection unit 150. The relationship information acquiring unit 160 acquires information (relationship information) of the relationship between the operation and the event by receiving an input from the worker who has performed the operation. Specifically, the relationship information acquisition unit 160 presents a question asking the relationship between the action and the event, and requests an answer by the worker who has performed the action. Then, for the presented question, an answer by the worker who has performed the operation is accepted. Therefore, the relationship information acquisition unit 160 is an example of a question presenting unit, and an example of an inquiry unit. And the relationship information acquisition part 160 is an example of an answer reception means. To present a question and receive an answer, for example, display a question screen having a question display unit displaying a question sentence and an answer input unit for inputting an answer on the display means, and accept the answer input to the answer input unit. It is done by Further, the relationship information acquisition unit 160 is an example of a request unit. Here, for example, questions prepared by the relationship information acquisition unit 160 may be prepared in advance, and a question statement corresponding to an action or a detected event for which the relationship is to be asked should be selected and presented. It is done by If the motion or event contains a variable (for example, when specifying the date and time, or when the number of occurrences of the event is the detection condition), those variables are acquired from the information where the event occurred and inserted in the question sentence Be done.

  Further, the relationship information acquisition unit 160 identifies an operation that has caused some influence based on the operation information managed by the operation information management unit 110 and the influence information managed by the influence information management unit 120. Here, when an action performed by the worker causes some influence, the action of the worker performed to cause the influence is referred to as the action of the worker. That is, the relationship information acquisition unit 160 specifies an action that has caused some influence among the operations managed by the operation information management unit 110. Then, the relation information acquisition unit 160 performs the above-described process for acquiring relation information regarding the action that is the specified action. Therefore, the motion information management unit 110, the influence information management unit 120, and the relationship information acquisition unit 160 are an example of an action identification unit.

  The identification of the operation causing the influence is performed, for example, as follows. Based on the operation information managed by the operation information management unit 110 and the influence information managed by the influence information management unit 120, rules for determining that an influence has occurred based on a certain operation are set in advance. For example, the relationship between the date and time when the operation specified by the operation information managed by the operation information management unit 110 was performed and the date and time when the influence identified by the influence information managed by the influence information management unit 120 occurred In the case of the above, a rule may be set to determine that such an influence has occurred based on such an operation. Further, when the target of the movement and the target which has the influence have a specific relationship, a rule may be set to determine that the influence has occurred based on the movement. Then, based on this rule, when it is determined that there is an influence generated based on a certain action, the relation information acquisition unit 160 acquires relation information regarding the action as an action that causes some influence. Do.

  Further, the relationship information acquiring unit 160 selects an event to be inquired about and makes a question regarding the operation identified as the action that causes some influence as described above. The selection of an event to be inquired is, for example, a rule in which such an event is to be inquired if there is a specific relationship between the date and time when the action specified as the action was performed and the date and time when the event occurred. Set up is done.

  The second information holding unit 170 is a holding unit that holds information on the relationship between the operation and the event acquired by the relationship information acquiring unit 160. That is, the second information holding unit 170 is an example of the second holding unit.

<Hardware configuration of computer>
FIG. 5 is a diagram showing an example of a hardware configuration of a computer used as the behavior information processing apparatus 100. As shown in FIG. The computer 200 shown in FIG. 5 includes a central processing unit (CPU) 201 which is a calculation unit, and a main storage unit (main memory) 202 which is a storage unit and an external storage unit 203. The CPU 201 loads a program stored in the external storage device 203 into the main storage device 202 and executes the program. As the main storage device 202, for example, a random access memory (RAM) is used. As the external storage device 203, for example, a magnetic disk drive, a solid state drive (SSD), or the like is used. Further, the computer 200 includes a display mechanism 204 for performing display output on the display device (display) 210, and an input device 205 on which an input operation by the operator of the computer 200 is performed. As the input device 205, for example, a keyboard, a mouse or the like is used. The computer 200 also comprises a network interface 206 for connecting to a network. The configuration of the computer 200 shown in FIG. 5 is merely an example, and the computer used in the present embodiment is not limited to the configuration example of FIG. For example, a non-volatile memory such as a flash memory or a ROM (Read Only Memory) may be provided as a storage device.

  In the behavior information processing apparatus 100 shown in FIG. 1, the operation information management unit 110 and the influence information management unit 120 are realized by, for example, the program-controlled CPU 201 and storage devices such as the main storage device 202 and the external storage device 203. Ru. The information collecting unit 130 is realized by, for example, the program-controlled CPU 201 and the network interface 206. The first information storage unit 140 and the second information storage unit 170 are realized by storage devices such as the main storage device 202 and the external storage device 203, for example. The feature detection unit 150 is realized by, for example, the CPU 201 under program control. The relationship information acquisition unit 160 is realized by, for example, the program-controlled CPU 201, the display mechanism 204 and the display device 210, and the input device 205. The question screen generated by the relationship information acquisition unit 160 is displayed on the display device 210 by the display mechanism 204. Then, when the operator operates the input device 205, an answer to the question shown on the question screen is input. In addition, in the operation information management unit 110 and the influence information management unit 120, when the operator himself / herself inputs the operation information and the influence information, for example, the input device 205 shown in FIG. 5 is used as the input means.

<Operation of Behavior Information Processing Device>
FIG. 6 is a flowchart showing the operation of the behavior information processing apparatus 100. The behavior information processing apparatus 100 causes the motion information management unit 110 to acquire and manage motion information (S601). Further, the behavior information processing apparatus 100 causes the influence information management unit 120 to acquire and manage the influence information (S602). Next, the relationship information acquisition unit 160 of the behavior information processing apparatus 100 performs the action of the worker based on the operation information held in the operation information management unit 110 and the influence information held in the influence information management unit 120. Are identified (S603). Further, the information collection unit 130 acquires information related to the object and the work (S604).

  Next, the feature detection unit 150 of the behavior information processing apparatus 100 detects information in which an event having a specific feature has appeared (S605). Then, the relationship information acquisition unit 160 generates and presents a question regarding the relationship between the operation identified as the action in S603 and the event having the specific feature in the information detected in S605 (S606). Next, when the relationship information acquisition unit 160 receives an input of a response from the worker who has performed the operation specified as the action in S603 (S607), the operation that became the query target according to the input response ( Behavior information and influence information, and information representing the relationship based on the answer are held in the second information holding unit 170 (S 608).

<First application example>
Next, application examples of the present embodiment will be described. Here, a process of detecting an action for facility maintenance will be described with a person in charge of maintenance of the facility facility as a worker. It is assumed that a maintenance worker who is a skilled worker (hereinafter, a worker) performs an operation of injecting 5 ml (milliliter) of oil into one of the cylinders constituting the work space facility. The identification information of the cylinder into which the oil is injected is # 22. Also, it is assumed that the injection of oil into cylinder # 22 was performed on June 2, 2017. Moreover, cylinder # 22 shall be provided in the installation installed in the work area "section X."

  Here, it is assumed that the worker is aware of the operation of injecting oil into the cylinder # 22 and the operation information is input by the operation of the worker himself and managed by the operation information management unit 110. Further, the operator recognizes that the operation of the cylinder # 22 becomes smooth as the influence of the injection of the oil, and the influence information is inputted by the operation of the operator himself and managed by the influence information management unit 120. Shall be Based on the operation information and the influence information, the relation information acquisition unit 160 sets the operation in which the oil is injected into the cylinder # 22 as the action of the acquisition target of the relation information.

  On the other hand, the information collecting unit 130 collects the object attribute information, the inspection value information, the work related information, and the worker operation information, and stores the information in the first information holding unit 140. Then, the feature detection unit 150 refers to the information in the vicinity of June 2 in which the action of injecting the oil into the cylinder # 22 by the worker is performed from the information held in the first information holding unit 140, Detect certain events. Here, referring to FIG. 2, it can be seen that six months have passed from the date of manufacture and two months have passed from the final inspection date as events relating to cylinder # 22. Also, referring to FIG. 3, it can be seen that the operation noise of the cylinder # 22 has increased by about 35% as compared to the previous day. Further, referring to FIG. 4, in the section X, the room temperature is 21 degrees at 11 o'clock to 12 o'clock, 22 degrees at 12 o'clock to 13 o'clock, 23 degrees at 13 o'clock to 14 o'clock and 14 o'clock to 15 o'clock . In addition, in the section X, it is understood that the worker is “Yamada” from 11 o'clock to 13 o'clock, and the worker is “Tanaka” from 13 o'clock to 15:00. Therefore, for example, assuming that the time when the action of injecting oil into cylinder # 22 is performed is 13:33, the room temperature is 23 degrees, and the worker is "Tanaka."

The relationship information acquisition unit 160 creates a question sentence for inquiring of the operator on the question screen regarding the event having the feature detected by the feature detection unit 150 based on a predetermined rule. Specifically, prepare a fixed-form sentence, “Who injected the oil from ○○○○?”, And in the “○○○○○” part, the contents based on the identified event insert. Here, for example, the following question sentence is created.
"Is it because oil was injected six months after production?"
"Is the oil injected because it has been two months since the last inspection of the cylinder?"
"Is it because the operating noise increased 35% compared to the previous day because of oil injection?"
"Is it because the room temperature was 23 degrees because the oil was injected?"
"Is the reason for injecting the oil because the room temperature was rising?"

The order of presenting the questions may be determined based on appropriate rules. For example, the following rules can be considered.
-Give priority to those with a large deviation from the average value of the values included in the information in the identified event.
-Give priority to those in which the value included in the information in the identified event exceeds a predetermined control value.
-Give priority to those with high degree of change in the value contained in the information in the identified event.
-Present in order from the event that occurred at the closest date and time to the date and time when the action for which the relationship information is to be acquired was performed.

  Furthermore, when there are a plurality of events detected by the feature detection unit 150, the presentation order of questions regarding the events may be determined according to the appearance degree of the features in each event. For example, as the event detected by the feature detection unit 150, an event P-1 "the sound has increased by 10 db or more" and an event P-2 "the air temperature has risen by 2 degrees or more" are set. Think. Then, it is assumed that the information collection unit 130 has acquired information I-1 indicating that the sound has increased by 30 db and information I-2 indicating that the temperature has risen 3 degrees. In this case, the event P-1 is detected by the information I-1, and the event P-2 is detected by the information I-2. Here, “30 db” of information I-1 with respect to “10 db or more”, which is a condition for detecting event P-1, and a condition for detecting event P-2 “2 The degree of expression “three degrees” of the information I-1 with respect to “degree or more” is compared. Then, if it is determined that the degree of expression of the information I-1 is larger than the degree of expression of the information I-2, then the question regarding the event P-1 is made first. In addition, the determination method and the determination standard of the size of the degree of appearance are specifically set according to the type of event, the environment of the site where the information collected by the information collection unit 130 is acquired, and the like.

  In addition, when detection of an event by the feature detection unit 150 is performed based on the worker operation information collected by the information collection unit 130, according to the certainty factor of the operation of the worker who has become the condition for detecting each event. The order of presentation of the questions regarding the event may be determined. For example, consider a case where worker operation information is obtained based on an image or a sensor value, and information on the degree of certainty of operation is added to the worker operation information. It is assumed that the information on each progress is based on a three-step evaluation with the maximum value when the operation is reliably performed. Then, regarding the motion detected as an event by the feature detection unit 150, the certainty factor that is a certain action B-1 is “3”, and the certainty factor that is another action B-2 is “2”. . In this case, a question regarding an event detected based on the determination that the operation B-1 having a high degree of certainty is performed is first performed, and the detection is performed based on the determination that the operation B-2 having a low degree of certainty is performed. Ask questions about events that will occur later.

  In addition, according to the kind of action which is the acquisition object of relationship information, the content of the detected event, etc., you may switch the rule which determines the presentation order of the question by a question screen. Moreover, as for the answer to the question sentence, in addition to the binary answer of the positive answer (for example, "Yes") and the negative answer (for example, "No"), the intermediate answer (for example, "○" , Answers with “△”, “×”), answers by the corresponding degree (for example, input some points out of 100 points), free answers by sentences (for example, input to the input field by text), etc. You may accept in various forms. Also, the number of questions presented may be controlled according to the content of the accepted answer.

  FIG. 7 is a diagram showing an example of the question screen. A question display unit 162 and an answer input unit 163 are provided in the question screen 161 shown in FIG. 7. The question display unit 162 displays a question sentence selected and created as described above. In the example shown in FIG. 7, the question sentence "Is the oil injected six months after production of the cylinder?" Is displayed. Further, in the example illustrated in FIG. 7, when the selection button 163 a for selecting one of “Yes”, “No”, and “Others” as an answer and “Others” are selected in the answer input unit 163. And an input field 163b for receiving an input of text. When the worker selects “Other” as the answer, the operator can input an answer to the question displayed on the question display unit 162 as a text by using the input field 163 b. As an example, in response to the question displayed in the question display section 162 of FIG. 7, the fact that "the cylinder has been six months after production" is the direct cause of the current "injected oil" action. However, if it is a matter to be taken into consideration in determining whether or not to inject the oil, a sentence having such content may be input in the input field 163b.

  FIG. 8 is a diagram showing another example of the question screen. In the question screen 161 shown in FIG. 8, the question display unit 162 is the same as the question display unit 162 of the question screen 161 shown in FIG. 7. The answer input unit 163 includes only an input field for receiving text input (described as “entry field” in FIG. 8). When the question screen 161 configured in this way is used, the relationship information acquisition unit 160 is an example of an input to the input field 163 b in the question screen 161 shown in FIG. 7 as an answer (relationship information) to the question. You can get answers that contain more specific content and detailed information. Further, answers such as “Yes” and “No” may be input to the answer input unit 163 shown in FIG. 8 by text input.

  When the relationship information acquisition unit 160 receives an answer, the operation information of the operation (here, the operation of injecting oil into the cylinder) specified as the activity and the influence of this operation (here, the operation of the cylinder becomes smooth) And the relationship information based on the reply of the worker (here, for example, an event that the operation sound is increased by 35% compared to the previous day) is stored in the second information storage unit 170.

<Second application example>
Next, another application example of the present embodiment will be described. Here, a process of detecting an action when a worker works on a product which is an object will be described. It is assumed that skilled workers and non-skilled workers perform different operations in the production line. Specifically, a skilled worker may change the standing position when performing work on a specific product.

  Here, it is assumed that an image obtained by photographing the work situation of the work place with a camera is analyzed by an external server (image analysis server), and an operation of a skilled worker not performed by a non-skilled worker is detected. Information of the detected operation is sent from the external server to the behavior information processing apparatus 100 and managed by the operation information management unit 110. In addition, the former by the manager of the work site with work by a skilled worker (worker who changes the standing position during work) and work by a non-skilled worker (worker who does not change the standing position during work) It is assumed that the defect emission rate is lower. This information is input as influence information by the administrator and managed by the influence information management unit 120. Based on the operation information and the influence information, the relationship information acquisition unit 160 sets the operation of changing the standing position when performing the work on the specific product as the action of the acquisition object of the relationship information.

  On the other hand, the information collecting unit 130 collects the object attribute information, the inspection value information, the work related information, and the worker operation information, and stores the information in the first information holding unit 140. The feature detection unit 150 refers to the information stored in the first information storage unit 140 and refers to information in the vicinity of the date and time when the worker took an action to change the standing position, and detect a characteristic event. Here, when the action of changing the standing position is frequently performed, there are a plurality of products which are targets of the work performed by changing the standing position. Then, for each product, there exist object attribute information, inspection value information, work related information, and worker operation information when work on the product is performed. Therefore, information classified as common types of information may be extracted and collected from information related to work for each product. Classification of information is performed in advance, for example, based on a specific rule.

The relationship information acquisition unit 160 creates a question sentence for inquiring of the worker regarding an event having a feature detected by the feature detection unit 150 based on a predetermined rule. The creation of the question sentence is the same as in the first application example. Specifically, for example, the following question sentence is created.
"Is there any relation between changing the standing position and the product installation angle on the conveyor of all target products being within the range of 45 degrees to 60 degrees?"
"Are 17 products (85%) out of 20 products whose standing position has been changed, are there any related products that were out of line with tightening torque value NG in the previous process?"

  Also in the second application example, for example, the order of presentation of question sentences may be controlled such that question sentences relating to events commonly occurring in work on more products are preferentially presented. When the relationship information acquisition unit 160 receives an answer, the operation information of the operation (here, the operation of changing the standing position) specified as the activity, and the influence of this operation (here, the defect emission rate decreases). ) And relationship information based on the worker's response (here, for example, an event that the product has been lined out with a tightening torque value NG in the previous step, and the product installation angle on the conveyor of the target product Is stored in the second information storage unit 170).

  As described above, in the present embodiment, among the actions of the worker, the action causing the influence is identified, and the characteristic event is detected based on various information collected in relation to the object or the work. Ask the worker the relationship between the worker's behavior and the detected event. Then, the reply by the worker is stored as knowledge about the action. By presenting the relationship between the worker's behavior and the event as a question, knowledge that the worker himself has verbalized and does not recognize (so-called tacit knowledge) is revealed, and this can be implemented as a method that can be implemented even for non-skilled workers. Help to extract.

  As mentioned above, although the embodiment of the present invention was described, the technical scope of the present invention is not limited to the above-mentioned embodiment. Various alternatives and modifications without departing from the scope of the technical idea of the present invention are included in the present invention. For example, the information of the variable collected by the information collection unit 130 is specifically set according to the type and use of the site system to be supported according to the present embodiment, the content of work, and the like. Also, the specific presentation method of the question is not limited to the above-described method. Further, in the second application example described above, the motion of a skilled worker not performed by a non-skilled worker is detected by analyzing an image obtained by photographing with a camera. The operator's action may be detected by specifying the standing position or posture of the worker, the tool possessed by the worker, the equipment to be handled, etc. from the sensor value acquired by the sensor.

100 Behavior information processing apparatus 110 Operation information management unit 120 Influence information management unit 130 Information collection unit 140 First information storage unit 150 Feature detection unit 160 Relationship information acquisition unit 170 ... 2nd information holding unit

Claims (8)

Collection means for collecting information on multiple variables in the work environment;
First holding means for holding information collected by the collecting means;
Action specifying means for specifying a worker's action in the work environment;
A detection unit that detects, from the information held by the first holding unit, information in which an event having a predetermined feature according to the type of the information has appeared;
Request means for requesting input of information on the relationship between the action specified by the action specifying means and the event of the information detected by the detection means;
Second holding means for holding information input in response to the request of the request means;
An action information processing apparatus comprising: The request means is
Question presenting means for generating and presenting a question;
Answer acceptance means for accepting input of an answer to the presented question;
The behavior information processing apparatus according to claim 1, comprising:   The behavior information processing apparatus according to claim 2, wherein the question presenting unit determines the type of the question and the presentation order according to the type of the behavior identified by the behavior identifying unit. The action identification means
Operation specifying means for specifying an operator's operation;
Impact identification means for identifying an event that is determined to be an impact of an operation;
The behavior information processing apparatus according to claim 1, comprising:   5. The action information according to claim 4, wherein the action specifying unit receives an input of information on an action by the worker, analyzes the input information, and specifies contents of the action of the worker. Processing unit.   The behavior information processing apparatus according to claim 4, wherein the operation specifying means specifies the content of the operation of the worker based on data of the worker measured by a sensor provided in the work environment. .   The behavior information processing apparatus according to claim 4, wherein the operation specifying unit analyzes an image obtained by photographing the worker and specifies a content of the operation of the worker. With regard to the information of the variable acquired in the work environment, information presenting an event having a predetermined characteristic according to the type of information is presented, and the relationship with the worker's behavior in the work environment is given to the worker Inquiry means to inquire,
Response receiving means for receiving an input of a response to the query by the query means;
An action information processing apparatus comprising:

Images